Abstract:
Cerebral palsy is a group of motor dysfunction conditions caused by non-progressive brain damage and is one of the most common childhood-onset, lifelong disabilities in New Zealand. There are no known cures to cerebral palsy, and the only means of improving a patient’s quality of life is through effective therapy interventions. Conventional rehabilitation approaches typically involve large amounts of contact time with a therapist who assesses the motor function of a patient before developing a labour intensive rehabilitation programme. However, as the mean population age increases, more stress is being placed upon healthcare systems around the world, making therapy less accessible and more expensive. One of the solutions to this issue has been to integrate robotics into lower-limb gait rehabilitation therapies. Robotic gait training systems reduce the labour requirements of therapists and are predicted to become cheaper as the price of hardware decreases. However, most of the existing robotic systems are stationary, do not cater to the needs of children and lack the ability to assess patient motor function. To address the shortcomings of these existing systems a novel paediatric robotic overground gait trainer (PRO-GaiT) has been developed. The PRO-GaiT utilises a five-bar linkage mechanism to manipulate the foot of a patient via an end-effector. Additionally, an accompanying assessment algorithm has been developed which is designed to evaluate a patient’s motor function. This assessment algorithm is comprised of a kinematic model, a series of quantitative movement metrics and a musculoskeletal model. The accuracy of the PRO-GaiT’s kinematic measures have been validated against an optical motion capture system. The PRO-GaiT was found to accurately measure joint kinematics with root mean square errors of 5.63 and 4.26 degrees for knee and hip angles respectively. The kinematic data recorded during this experiment was then processed by the assessment algorithm. The kinematic model and quantitative movement metrics successfully characterised the motor function of each participant and the musculoskeletal model simulated in-depth muscle properties such as muscle fibre lengths, muscle power and muscle activations.